[openstack-dev] [keystone] External authentication plugins

2013-12-19 Thread Brant Knudson
We've got to figure out what external authentication plugins we're going to
provide in Keystone.

This is something that you'd think wouldn't be complicated, but somehow
it's gotten that way.

Since we've made mistakes in the past, let's try to be careful this time
and come up with what plugins are required, and make sure those are
implemented.

To this end, I've opened a blueprint:
https://blueprints.launchpad.net/keystone/+spec/external-auth-plugins

What I need is,
a) Make sure the background info is correct. It documents the plugin
behavior that we've provided in the past and how they work. Keystone must
continue to support these for backwards-compatibility.

b) Figure out if there are new plugins that we need to provide. For
example, we don't have a V3 plugin that works like V2 authentication; we
don't have a V3 plugin that works like Grizzly did.

Thanks! - Brant
___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


[openstack-dev] [keystone] External authentication

2013-11-14 Thread Álvaro López García
Hi all,

During the review of [1] I had a look at the tests that are related
with external authentication (i.e. the usage of REMOTE_USER) in
Keystone and I realised that there is a bunch of them that are setting 
external as one of the authentication methods. However, in
keystone.auth.controllers there is an explicit call to the external
methods whenever REMOTE_USER is set [2].

Should we call the external authentication only when external is set
(i.e. in [3]) regardless of the REMOTE_USER presence in the context?

[1] https://review.openstack.org/#/c/50362/
[2] 
https://github.com/openstack/keystone/blob/master/keystone/auth/controllers.py#L335
[3] 
https://github.com/openstack/keystone/blob/master/keystone/auth/controllers.py#L342
-- 
Álvaro López García  al...@ifca.unican.es
Instituto de Física de Cantabria http://alvarolopez.github.io
Ed. Juan Jordá, Campus UC  tel: (+34) 942 200 969
Avda. de los Castros s/n
39005 Santander (SPAIN)
_
Unix never says `please.' -- Rob Pike

___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev


Re: [openstack-dev] [keystone] External authentication

2013-11-14 Thread Adam Young

On 11/14/2013 10:52 AM, Álvaro López García wrote:

Hi all,

During the review of [1] I had a look at the tests that are related
with external authentication (i.e. the usage of REMOTE_USER) in
Keystone and I realised that there is a bunch of them that are setting
external as one of the authentication methods. However, in
keystone.auth.controllers there is an explicit call to the external
methods whenever REMOTE_USER is set [2].

Should we call the external authentication only when external is set
(i.e. in [3]) regardless of the REMOTE_USER presence in the context?
I'd like to.  We made a decision to make the user explicitly enable 
External authentication in the config, but there is no reason that it 
would have to extend to the request body itself.
In theory we could do token creation request with no Body at all, the 
same way we do role assignments:


To create a project scoped token
PUT /auth/tokens/domain/domid/projectprojectid

And to create a domain token
PUT /auth/tokens/domain/domid


Would work very well with Basic-Auth or other External formats. Then the 
Body would only have to contain any mitigating factors, like a shorter 
expiry or reduced set of roles.







___
OpenStack-dev mailing list
OpenStack-dev@lists.openstack.org
http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev